How to use the video to audio
- Drop a video — or an audio file you want to convert.
- Pick a sample rate: 16 kHz mono is what speech recognition wants.
- Click Extract audio and wait for the decode.
- Download the WAV.
Video files are containers: an H.264 video track plus an AAC audio track. Extracting the audio means re-encoding it into a format you can use, and the useful choice is almost always uncompressed 16-bit WAV — every editor, transcriber and DAW reads it, and there is nothing to lose in a second lossy generation.
Choosing a sample rate
- 16 kHz mono — speech recognition, voice notes, podcast transcripts. Smaller, and it matches what Whisper-class models expect.
- 44.1 kHz — music, or the original rate for CD-style audio.
- 48 kHz — anything headed for video editing.
Converting from stereo to mono sums the channels, which is right for speech. Expect roughly 10 MB per minute at 44.1 kHz stereo, about 2 MB per minute at 16 kHz mono.
Why not MP3
MP3 encoding in a browser needs a large JavaScript encoder, and it would throw away detail for no benefit if the next step is transcription or editing. WAV is the honest default; convert to MP3 afterwards if the file has to be small.
Frequently asked questions
Does it work for long videos?
Decoding happens in memory, so a few hundred MB is the practical limit on a normal machine. Very long files are better handled with ffmpeg on the command line.
Is the audio quality preserved?
The output is WAV, so there is no further loss beyond the lossy source. It cannot restore detail the original AAC track already discarded.
Why does decoding fail on some files?
Some containers use codecs the browser cannot decode, such as AC-3 or some HEVC variants. Convert the source to MP4/M4A first.
Is anything uploaded?
No — the file is decoded locally with the Web Audio API.